D1 Part 3 leftovers: remove 8 residual emoji + Python emoji gate

- The D0 grep gate false-negatived multibyte emoji on Windows (its zero was wrong). scripts/emoji_gate.py scans by Unicode codepoint; it found 8 survivors: backups.html (📁→file-text icon, 🔄 restore-info text), debug.html (🔄→'fut', 🔒→'titkosított'), deploy.html (📦→upload icon, ★ default-marker → '(alapértelmezett)'), storage.html (📦→upload icon, 📦 in the migrate-target option → plain text). All → sprite icons or plain words. - Gates: scripts/emoji_gate.py = 0; new Go TestNoEmojiInTemplates codepoint scan keeps it enforced (allowlist: ✓✗✔✘•●○■▶ monochrome text marks). go build/vet/test ./... green (18 pkgs).
2026-07-02 19:19:23 +02:00
parent cb6f04c8fc
commit 622d9328f8
6 changed files with 115 additions and 8 deletions
@@ -0,0 +1,69 @@
+# -*- coding: utf-8 -*-
+"""D1 §10 Python emoji gate — Windows grep silently fails to match multibyte emoji (the D0
+grep-gate zero was a false negative). This scans templates by Unicode codepoint.
+
+Run from controller/:  python scripts/emoji_gate.py
+Exit 1 if any emoji/pictograph remains in web+setup templates.
+"""
+import io, os, sys, unicodedata
+
+ROOTS = [
+    os.path.join("internal", "web", "templates"),
+    os.path.join("internal", "setup", "templates"),
+]
+
+# Codepoint ranges that count as "emoji / pictographs / dingbats" for UI-copy purposes.
+# Deliberately does NOT flag Hungarian letters, typographic quotes/dashes, the middle dot (·),
+# arrows used as affordances (→ ↗ ↻ ↑ ↓), the multiplication sign (×), or box-drawing.
+def is_emoji(ch):
+    o = ord(ch)
+    ranges = [
+        (0x1F300, 0x1FAFF),  # Misc symbols & pictographs, emoticons, transport, supplemental, symbols-ext
+        (0x2600, 0x26FF),    # Misc symbols (☀ ⚙ ⚠ ☁ …)
+        (0x2700, 0x27BF),    # Dingbats (✅ ✂ ✈ ✏ ✓? no — see allow)
+        (0x1F000, 0x1F0FF),  # Mahjong/dominoes/cards
+        (0xFE00, 0xFE0F),    # Variation selectors (emoji presentation)
+        (0x1F1E6, 0x1F1FF),  # Regional indicators
+    ]
+    if any(a <= o <= b for a, b in ranges):
+        return True
+    return False
+
+# Dingbat codepoints that are legitimate UI glyphs (checkmarks/crosses used as plain text marks,
+# not emoji). We keep these OUT of the ban — they render as monochrome text, not color emoji.
+ALLOW = set("✓✗✔✘•●○■▶")  # ✓ ✗ ✔ ✘ • ● ○ ■ ▶
+
+
+def scan(path):
+    hits = []
+    for lineno, line in enumerate(io.open(path, encoding="utf-8"), 1):
+        for ch in line:
+            if ch in ALLOW:
+                continue
+            if is_emoji(ch):
+                try:
+                    name = unicodedata.name(ch)
+                except ValueError:
+                    name = "U+%04X" % ord(ch)
+                hits.append((lineno, ch, name))
+    return hits
+
+
+def main():
+    total = 0
+    for root in ROOTS:
+        for fn in sorted(os.listdir(root)):
+            if not fn.endswith(".html"):
+                continue
+            path = os.path.join(root, fn)
+            for lineno, ch, name in scan(path):
+                total += 1
+                print("%s:%d  %s  %s" % (fn, lineno, ch, name))
+    if total:
+        print("EMOJI GATE FAILED: %d emoji found" % total)
+        sys.exit(1)
+    print("emoji gate OK — no emoji in web/setup templates")
+
+
+if __name__ == "__main__":
+    main()